丘陵山区田间道路智能农业装备的自主避障算法研究

Research on Autonomous Obstacle Avoidance Algorithm of Intelligent Agricultural Equipment on Field Roads

由于丘陵山区田间道路复杂的环境,智能农业装备难以实现高安全性的自主避障功能.为此,融合多源传感器信息,提出一种基于栅格地图优化方法和改进A星算法的自主行驶避障方案.运用优化的Bresenham's算法和滚动更新原理投射障碍物位置信息,使栅格地图更具有实时性和可靠性,形成一种动态栅格地图.改进的A星算法采用代价函数分别评估安全性、局部路径与全局路径的偏移度.最后,利用ROS系统进行离线仿真和实时真实道路试验.仿真和试验结果表明:规划路径的曲率波动范围较小,提出的栅格地图优化方法和改进的A星算法满足丘陵山区田间道路低速智能农业装备避障的安全要求.

Due to the complex environment in field roads,it is difficult for intelligent agricultural equipment to achieve high-safety autonomous obstacle avoidance.Therefore,by fusing multi-source sensor information,this research proposes an autonomous obstacle avoidance method based on grid map optimization and an improved A*algorithm.The optimized Bresenham s algorithm and rolling update design make the grid map more real-time and reliable,which forms a kind of dynamic grid map.The improved A*algorithm uses cost functions to evaluate the safety and the deviation between the local path and the global path,respectively.Finally,use the ROS system for offline simulation and real-time real road tests.Simulation and real-time test results show that the fluctuation of curvature of the planned path is small.The proposed grid map optimization method and improved A*algorithm meet the safety requirements of low-speed intelligent agricultural equipment obstacle avoidance on field roads.